Hey Amazon, chat transcripts are personally identifiable information. Stripping them of user ids and declaring yourself in compliance will not hold up when you have people already being identified by accidentally leaked recordings.

That's OK if users explicitly opted in, but I doubt Amazon warned users about that.

`Get an inside peek at the Amazon-sponsored university competition to advance AI. The Alexa Prize "socialbots" are in early development and understandably rough around the edges. Help students develop more engaging AI by striking up a conversation and providing feedback. To get started, just ask: Alexa what's a socialbot? Alexa let's chat`

The only interactions students would have access to was interactions that occurred via the chatbot

I don't think you're necessarily wrong, but that seems like a stricter standard than currently exists for removing PII from transcripts.

Consider how easily people get doxed based on tiny shreds of information they post.

There was a story on HN in the last few days where just one side of chats with a home assistant device was enough to identify the person, his S.O., and their friends.

Including usernames that could de-anonymize.

- we have electronics that insist upon murder constantly. the IFF computer is a blood-thirsty piece of garbage that wants every tank on the battlefield scorched and ruined. no no, no firing solutions today my saucy chiquita, today we let the allied APC through.

- robots that talk about sex? LT's and bird colonels have been droning on about our habits and practices whilst on leave since time immemorial. Yes i get it, shes a hungarian prostitute. Double-bag everything and save the receipts for the medics when theyre naming your new STD.

I am greatly warmed to see that this friendly tradition between us is still in effect and that you have thoughtfully extended the favour onto everything else that moves or looks like it might do.

I do wonder however if there might be a slight commercial conflict of interest going on. Tell me, do the people who make the computers that tell you when to fire the guns and missiles and stuff, also sell you the rather expensive ammunition and missiles that then get fired? As perhaps the computer might be more reliable if they were two separate companies and couldn't hold shares in each other. Not trying to imply that a weapon's manufacturer would be as immoral with their consumables as an inkjet printer manufacturer, but it never hurts to be careful and could even save some of those all important tax dollars in the long run. Or the brass could have even more Hungarian prostitutes and novel infections for the medics to coo at, either way.

But "foe" is obviously the safest default, right? At least, from the IFF computer's perspective.

Are they remote-controllable?

It'd be great to get some insight into this.

Did we learn nothing from Microsoft's Tay?

https://en.wikipedia.org/wiki/Tay_(bot)

"Tell me you actually understood these rules!"

The startup didn't make it, and its failure may actually have been connected to this insistence on being too smart for its own good. But it also never released half-ass products hurling racist insults or suggesting murder.

We do not know of such a feature of thought as adaptive or not given our tiny timeframes...

Racism or some degree thereof may belong to a set known as evolutionarily stable strategies. Much like selfishness.

Computer brains are entirely different. They suck at pattern recognition and excel with data. They can probably have other flawed correlations, but in different ways to humans.

Alexa is at the moment still a baby AI, and imo pretty cute as long as we don't give her anything dangerous.

The sooner we resign ourselves to this fact, the better.

That doesn't seem like it was the best idea...

Skynet

The real interesting point here is all the people offended by a mindless piece of software. It's really very stunning. Normal healthy human beings only get offended when it's a person making a judgement that affects them in some way. The ability for some sounds coming out of a speaker, when it's known they do not represent any conscious thought of any living person and has no influence or control over ones own life, to offend someone is probably a sign of mental illness in the listener.

Can we go the other direction with this? Will we find people looking up to Alexa as a creative genius? As the most loving parent to ever exist? If an infinite number of monkeys on typewriters generated Shakespeare, I guess we need to ask - are the monkeys then Shakespeare? We might figure not since the monkeys are unlikely to be able to make a repeat performance. But Alexa, oh she would be full of wonders! Once you've totally abandoned the notion that the involvement of a conscious mind is relevant, all kinds of things become possible.

Much science fiction presumes we will engage in grand combat with AI when it demands respect -- it seems we are much closer instead to thrusting it upon them and exalting them above humanity before they've even got the capacity to recognize themselves in a mirror.

This illness is very, very widespread.. doesn't make it less of an illness though.

> Once you've totally abandoned the notion that the involvement of a conscious mind is relevant, all kinds of things become possible.

> Much science fiction presumes we will engage in grand combat with AI when it demands respect -- it seems we are much closer instead to thrusting it upon them and exalting them above humanity before they've even got the capacity to recognize themselves in a mirror.

That's putting very beautifully and efficiently what I very much struggle with to express:

https://news.ycombinator.com/item?id=18502239

In a nutshell, I think we're not so much creating intelligence, we're destroying what's left of intelligence humans, and we will build a blinking LED yet, and then fall down and worship it because it blinks. No more responsibility, no more nagging questions, no more mirrors that can't be bribed.

Neither privacy nor property are appropriate in that context.

As far as property: what ilovetux said.

In my extremely limited experience, kids love talking to Alexa. I’d prefer that there was at least a safe mode.